Hi Tobias,
This unifies the two locus to location_t conversion functions, preparing
for some changes I want to do later.
In principle, I had the patch this morning; however, the assert is now
exercised more often than before - and it triggered rather unexpected
when running the testsuite.
Turned out that partially overriding a derived-type spec with a string
length is not really a good idea ...
Otherwise, it is a rather simple patch.
Nonetheless, are there any comments, suggestions, concerns?
Tobias
PS: I have added the assert to gfc_resolve_code, gfc_resolve_expr and
resolve_symbol without additional fails; this does not guarantee that
there isn't any issue with other locations the code, but a least it
makes it less likely. As there was no good reason to include that
assert, I have done so for the attached patch.
The patch looks good to me, and fixes a long-time issue (not serious,
but mildly annoying). Just one nit: I think it would be
enough to add a gcc_checking_assert(), a normal user using a release
version is not likely to profit from the assert. This is the
same approach that we took in frontend-passes.cc.
Best regards
Thomas
